3rd Shift doctrine

01Why this matters

1. When PJ asks "can we automate X?": walk the 5 criteria. If any answer is no → keep manual + propose the missing guardrail. 2. When a daily ritual surfaces: evaluate if it's mechanical enough to cron. The friction-compounds doctrine + this doctrine work as a pair — friction-compounds says automate recurring work, this says only WHERE guardrails can catch the failure modes. 3. When PJ goes AFK and the riff doctrine fires: "Project Clean Riff" mode IS day-shift autonomous (Claude as PJ-proxy). 3rd shift is fully autonomous (no PJ proxy). Different scopes. 4. When new failure modes surface in production: propose the guardrail BEFORE proposing the automation. The guardrail compounds; the automation is one-shot.

02How to apply

Read the doctrine. Apply it on the next ship cycle.

03The closer

PJ-validated 2026-05-06 framework for when to expand SideGuy's autonomous overnight cron layer (3rd shift) vs when to keep work in PJ's day-shift hands. Mechanical + guardrail-tested = automate. Voice + judgment = stay manual. Connects to the intelligent-expansion-guardrails doctrine.

04Text PJ — no Calendly, no meeting